Feeds

back to article NASA confirms manned mission to 10 Petaflops

Well, well, well. It would seem that the 20,000-core supercomputer announced yesterday by NASA will just be the first course in an ongoing relationship between the space folk, SGI and Intel. The three organizations have revealed a project dubbed Pleiades that will see them build Petaflop-class machines in the coming years. NASA …

COMMENTS

This topic is closed for new posts.
Silver badge
Coat

Necessary? To the moon?

In the 60's when the USA first went to the moon, the computers weren't much more than what is in a cell phone (really much less). Why all the compute power? Hasn't anyone learned how to make algorithms efficient.

Oh, I forgot, they're using Windows. I'll get my coat.

0
0
E

"...fuel the manned missions to the moon..."

How sad that we went to the moon using slide rules and computers far, far less powerful than my HP RPN calculator. Now we require monstrous super computers to repeat something done forty years ago.

On the bright side, maybe with a few petaflops and some virtualization, NASA will be able to run *two* instances of Crysis with full graphics on.

0
0
Black Helicopters

Better fakes

at least now the fake video and photos from the next moon mission and mars trip will be harder to spot

0
0
Silver badge

Well, they haved a budget to use...

I doubt it is really necessary to use that much power for the mission.

But well, how else are they going to justify spending so much of the taxpayer's money?

0
0
Alien

This is NASA we're talking about...

...NASA don't do things efficiently. It took them 30 years to realise the Shuttle was the "wrong way around*". Well... that's not true. They *KNEW* it was the wrong way around, but it was a good excuse to pour time, effort and money into some really cool engineering.

*By "wrong way around" I mean that it *should* be the first stage that's re-usable, with an expendable payload delivery system (a lá ESA's ATV). Much more efficient, especially in terms of cost of repair/maintenance. It's just nowhere near as cool.

0
0
Coat

Depends...

..on the graphics card.

"On the bright side, maybe with a few petaflops and some virtualization, NASA will be able to run *two* instances of Crysis with full graphics on."

Mine's the on with 2,048 NVIDIA® GeForce™ 8800GT's and a blu-ray drive.

0
0
Stop

@Herby

Why do we always have to have a "I forgot, they're running Windows" comment?

Especially seeing as they're running SuSe Linux

0
0
Anonymous Coward

If only they had said Jupiter...

It would've been easier to swallow...

0
0
Paris Hilton

Better Fakes 2

"at least now the fake video and photos from the next moon mission and mars trip will be harder to spot"

I feel this is getting closer to the truth

0
0
Anonymous Coward

@E

"How sad that we went to the moon using slide rules and computers far, far less powerful than my HP RPN calculator. Now we require monstrous super computers to repeat something done forty years ago."

Hmmmm. You sure about that?

I am still incredulous that America keeps banging on about how they "apparantly" managed to get to the Moon in the 60's on 60's technology, but now NASA needs these HUGE computers to do it again. More technology comes out of my cat's arse when he has diarrhea.

Well, he's getting old, no teeth left, farts a lot etc...but I digress.

Personally I cannot wait for the day when the Chinese get up there, wander around for a bit and then say "Where did you said you landed? Just here? Just on this spot riiiiigggghhhtttt here? You bunch of lying twats"

God, I hope I'm alive. And that the Yanks haven't either invaded or nuked China before they get there.

0
0
Flame

California?

I never understand why these guys keep insisting on building their supercomputers in the hottest parts of the world. Why not build it in Alaska and save a fortune in air conditioning?

0
0
Anonymous Coward

Re: California?

Because then the sea levels would rise way before global warming becomes deadly.

0
0
Gates Halo

Extra speed

So we should be seeing Petaflops in my laptop in about 10 years?

0
0
Joke

If NASA's new supercomputer is called "Pleiades"...

Is DARPA's next supercomputer going to be called "The Wizard of Wor"?!

Would the Department of Sanitation's supercomputer be called "Pong"?!

Will the USDA's supercomputer be called "Track & Field"?!

If Bush and Cheney shared a computer, would it be called "Donkey Kong"?!

0
0
Linux

@herby

No , they clearly stated that it will be running linux. I guess they need so many cores to run all the various releases with different patchlevels and all the different strains of 'looserix' (sometimes also called 'loonix' ). Especially since all their software needs to be recompiled every time some 'open sores' / 'broken source' tweak breaks another thing ..

pengiuns.. should have left them at the south pole.

0
0

@Ian Ferguson

last time I checked mt view is not the hottest place in the worlds , neither is Edwards California which is were Edwards air force base is. Right now in Stockton ca its 75ºf Edward,ca its 80ºf Lost wages oops I mean Las Vegas ,NV its 85 ºf Tuscon AZ its 86ºf

0
0

@Plankmeister

'...NASA don't do things efficiently. It took them 30 years to realise the Shuttle was the "wrong way around*". Well... that's not true. They *KNEW* it was the wrong way around, but it was a good excuse to pour time, effort and money into some really cool engineering.'

Blaming NASA for the Shuttle is completely wrong. The Shuttle was a compromise between those in the Nixon administration which looking for a follow-on to Apollo, but for a fraction of the cost; and those in the USAF who were looking for a combo orbital bomber and space truck. By the time the Shuttle was approved, the Saturn line had already been closed and there was no alternative way of getting man into space without asking the Soviets.

The original Shuttle plan of the 1960s called for a reusable booster carrying a piggybacked orbiter, all liquid fuelled; to serve a giant space station akin to the one in 2001 which would be in charge of constructing the monstrous Mars missions. The recession of the late 1960s, rampant inflation and the war in Vietnam killed off the Mars mission then the space station. All that was left was the Shuttle, and that on a dramatically reduced budget.

USAF called the shots and dictated the size of the payload bay and the need for a winged glider that could return to base anywhere in the US after making a single polar orbit. Nixon and Congress killed the reusable component by hacking the budget so that the External Tank would be discarded and cheaper reusable solid-rocket boosters substituted for liquid-fuelled boosters.

NASA designed the Shuttle to fit its vastly diminished budget (and blew that as well). It was a botched job - a brilliant botch - but the failure of the Shuttle lies in politics not in the incredible engineering of the machines. It's worth pointing out that no other country has yet replicated many of the Shuttle's features and that the Soviet Buran orbiter was far from finished when it made its single, even more expensive flight.

0
0

Apollo, slide rules and petaflops

It's worth remembering that these machines won't be flown on the missions, just used to simulate parts of the flight equipment.

The Constellation programme is still much less lavishly funded as a share of GDP than Apollo ever was; it is also far more sophisticated, being partially reusable and intended for far longer duration flights.

There simply isn't the money even in America's budget to try the 'throw everything and see what sticks' approach of the 1960s. It's worth remembering the Saturn V that took man to the Moon wasn't the only program involved - there were also Ranger, Surveyor and Lunar Orbiter doing the reconaissance, Gemini proving most of the techniques needed to fly to the Moon and the Saturn IB launcher as an intermediate technology.

Apollo was rushed, fatally so with the fire of Apollo 1 and the flight of Apollo 8 around the Moon was a huge risk undertaken to avoid the humiliation of a Soviet Zond being the first manned mission to make that journey (Zond itself was canned when its booster cracked). The Saturn V was a beast to fly and nearly shook itself to pieces on early flights and there is always the near disaster of Apollo 13 in the back of NASA's mind. They're not going to take the same risks again.

If anyone is interested in this, I can't recommend 'In the Shadow of the Moon' and 'From the Earth to the Moon' highly enough. The first is a series of interviews with many of the Apollo astronauts, the second dramatic reconstructions of the Apollo missions.

0
0

Misconception

As a person who works an Ames and routinely uses super computers all over the world I'd like to point out a few misconceptions.

They commonly refer to these big machines as being used for a certain task but in reality 100's of jobs are running on them at any one time. The jobs can be from all different types of disciplines ranging from atmospheric science to rotorcraft aerodynamics to biological systems modeling. In addition many jobs are run by university students who are given computer time as part of their research on government contracts.

In reality, I'd say less than 5% of the time on the new computer will be spent supporting missions to space but that doesn't mean the machine will not be put to good use. It is just easier to sell the idea to funding sources by associating a sexy task with it.

For those gamers out there:

The machines have no graphics cards on them. All the data is transferred back to work stations for the post processing of results and visualization. The machines sole purpose is to crunch numbers as fast as possible.

To the Windows person:

Windows is far from ideal when most of the code is written in Fortran and C using MPI or OpenMP for parallelization.

0
0

Supercomputing and the moon

Computers were used extensively in the original Apollo program, but better computers and other new technology will allow us to do the same thing even better for a lot less money.

0
0
Black Helicopters

@Lukin

<<If Bush and Cheney shared a computer, would it be called "Donkey Kong"?!>>

They do - it is an etch-a-sketch.

Oh damn they spotted me. Gotta run!

0
0

@ George Schultz

Yes and Bush is still trying to find the on/off switch

0
0

on power on...

When the on button is pressed, do the lights dim for a few seconds?

20,000 CPUs. If 1 CPU takes up 3 cubic inches -- you have to count the fan on top of the CPU. 20,000 x 3 = 60,000 cubic inches or a box a little over 39 inches in length of just pure CPUs. A 1 gig memory stick is about 5 inches long. Assume 5 inches x 20,000 = 100,000 inches of memory -- or 8,333.33 feet of memory. I hope they have a long building to put this machine :) I can just imagine, where the heck do I plug in my USB device?

If Matlab (a software for mathematical calculations) costs $1,900 per CPU, the unexpected expense of the software will be 38,000,000 -- I hope they use Octave (the open source alternative).

I can see it now. 20,000 CPUs waiting for a 7200 RPM hard drive to spin up....UGH!!! Blocked on I/O again!

Imagine writing the code that accidently deadlocked 20,000 CPUs and getting your name in the Guiness Book of World Records!!! You could be fired and famous on the same day!!!

Imagine trying to find the bad (CPU, memory stick, etc) -- talk about a possible NIGHTMARE!!!

Imagine booting .... KERNEL PANIC....dialing Kernelpanic.org --- "Hello, is this kernelpanic.org? .... I have a problem I hope you can help me...."

When will this computing power reside in my handheld?

0
0
Happy

@James Moliere

1) There's no fan on top of a server CPU, they use a big heat sink and several redundant fans in the blade chassis.

2) Servers don't use 7,200rpm drives, they use 15,000rpm SAS drives.

3) These boxes use 8gb RAM modules, so divide your calculation by 8, then there's the fact that the memory is in parallel slots, much like your desktop machine, so it doesn't have to be an 8000 foot long room.

4) As for finding bad RAM, CPU etc, all the hardware is managed so a simple management console reports on the status of all hardware, giving details of the cabinet, shelf & blade that holds the dodgy component.

I have to agree about the power though, I always imagine that these machines should come with one of those Frankenstein style switches that can only be activated during a thunderstorm to give sufficient current to fire up 20,000 CPUs and a similar number of fast hard drives...

0
0

@James Moliere

20,000 cores, NOT 20,000 CPU's

0
0

Does this mean what I think it means?

"NASA reckons that such machines will be needed to fuel the manned missions to the moon"

as in... an actual admittance of... erm, well faking the whole thing in 69?

0
0

Tenzing Norgay

When ecodopes go on their penguin fondling cruises they expect not to die and to live in luxury. that surely should be the difference in the experience for those who return to the moon. Vast amounts of compute power is lifeboats and champagne.

0
0
Linux

@Ian Ferguson

Ames is located right on the edge of the San Francisco Bay (next door to SGI) and the weather tends to be on the chilly and breezy side. Average high temperatures in July are not much different from London. Hot days are rare.

Linux is the only option for scaling to a huge cluster without huge expense.

0
0
Joke

fuel the manned missions to the moon

Surely rocket fuel would be better?!?!

Anyway isn't Pleiades the Seven sisters constellation? so I guess there will be seven seperate machines in this cluster then?

0
0

Re: Necessary? To the moon?

"In the 60's when the USA first went to the moon, the computers weren't much more than what is in a cell phone (really much less). Why all the compute power? Hasn't anyone learned how to make algorithms efficient.

Oh, I forgot, they're using Windows. I'll get my coat."

The 20480 core machine is actually using Suse Linux. Read the article yesterday..

Not that I am saying Linux is a massive resource hog..

0
0
Alien

@ Anonymous Coward

"Personally I cannot wait for the day when the Chinese get up there, wander around for a bit and then say "Where did you said you landed? Just here?""

Maybe that's exactly why the US is in such a hurry to go back to the moon - they have to get up there and plant all the stuff that was supposed to be left behind in the 60's.

An alien... because I love conspiracies and cover up stories.

0
0
Anonymous Coward

@ Plankmeister

"...NASA don't do things efficiently. It took them 30 years to realise the Shuttle was the "wrong way around*". Well... that's not true. They *KNEW* it was the wrong way around, but it was a good excuse to pour time, effort and money into some really cool engineering.

*By "wrong way around" I mean that it *should* be the first stage that's re-usable, with an expendable payload delivery system (a lá ESA's ATV). Much more efficient, especially in terms of cost of repair/maintenance. It's just nowhere near as cool."

You mean that the boosters should be reusable and the shuttle not so..? Hmm real smart. Do they use disposable astronauts too..? And the orbiter is the most expensive bit of the shuttle stack.

Also, every bit of the shutttle is re-used except for the big orange external tank. The SRB's are recovered and used too.

0
0
This topic is closed for new posts.